Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user needs to estimate market size, understand market dynamics, or validate that a market opportunity is large enough to pursue.
.claude/skills/mkurman-market-research/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 81% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 32% | 0% |
| case-13 | ✓→✗ | ▼ Worse | 54% | 0% |
|---|---|---| | TAM | $X | $Y | Top-down + Bottom-up | High/Med/Low | | SAM | $X | $Y | Filtered from TAM | High/Med/Low | | SOM | $X | $Y | Penetration model | High/Med/Low |
Top-Down: Step-by-step calculation from industry totals to target segment. Show every filter and assumption applied.
Bottom-Up: Step-by-step calculation from unit economics up. Show: (number of target customers) x (expected conversion rate) x (annual contract value).
Reconciliation: Comparison of both approaches, explanation of any gaps, and reconciled estimate. If they diverge by more than 3x, investigate the assumptions driving the gap.
Key factors that could expand or contract the market -- technology shifts, regulatory changes, demographic trends, behavioral changes, and emerging adjacent segments.
| Assumption | Impact if Wrong | Confidence | Validation Method | |---|---|---|---| | Description | What changes in the estimate | High/Med/Low | How to test this |
Numbered list of what the market data means for product, pricing, and go-to-market decisions.
competitive-analysis -- Pair market sizing with competitive landscape analysis to understand both the size of the prize and how contested it is.prd-writing -- Use market segment data to ground the Market Segments section of a PRD in real numbers.roadmap-planning -- Use growth trend analysis to time roadmap investments. Build for fast-growing segments first.User: "Help me size the market for our developer productivity tool. We need the TAM/SAM/SOM for our Series A deck."
Good output excerpt: > TAM: $X]B -- Global developer tools market (source: industry report], includes IDEs, testing, CI/CD, monitoring, and productivity tools). > > SAM: $4.2B -- Code review and collaboration segment, filtered to teams of 10-500 developers at companies with >$5M revenue in North America and Europe. > > SOM: $85M -- 2% penetration of SAM over 4 years, based on current growth rate of 15% QoQ and average ACV of $18K. > > Bottom-up cross-check: 23,000 target companies x 12% expected conversion at maturity x $18K ACV = $49.7M.
User: "We're thinking about expanding from SMB to mid-market. Is that market big enough?"
Good output should size the mid-market segment separately, compare unit economics (higher ACV but longer sales cycle), estimate the investment required to serve the segment (enterprise features, sales team), and calculate whether the segment-level SOM justifies the investment within the planning horizon.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 38,045 | 25,359 | -33% | 1 | 1 | 0% | 6,237 | 5,459 | -12% | 0 | 0 | — |
case-02 | fail→fail | 35,739 | 26,774 | -25% | 1 | 1 | 0% | 6,233 | 5,625 | -10% | 0 | 0 | — |
case-03 | fail→fail | 37,830 | 38,084 | +1% | 1 | 1 | 0% | 6,219 | 7,393 | +19% | 0 | 0 | — |
case-04 | pass→pass | 15,452 | 14,535 | -6% | 1 | 1 | 0% | 2,745 | 3,696 | +35% | 0 | 0 | — |
case-05 | pass→pass | 19,563 | 23,373 | +19% | 1 | 1 | 0% | 3,245 | 4,934 | +52% | 0 | 0 | — |
case-06 | pass→pass | 12,368 | 15,452 | +25% | 1 | 1 | 0% | 1,943 | 3,625 | +87% | 0 | 0 | — |
case-07 | pass→pass | 12,807 | 28,796 | +125% | 1 | 1 | 0% | 2,062 | 3,868 | +88% | 0 | 0 | — |
case-08 | pass→pass | 14,570 | 17,598 | +21% | 1 | 1 | 0% | 2,401 | 4,131 | +72% | 0 | 0 | — |
case-09 | pass→pass | 19,441 | 30,539 | +57% | 1 | 1 | 0% | 3,565 | 6,130 | +72% | 0 | 0 | — |
case-10 | pass→pass | 14,108 | 14,380 | +2% | 1 | 1 | 0% | 2,607 | 3,917 | +50% | 0 | 0 | — |
case-11 | pass→pass | 8,246 | 5,685 | -31% | 1 | 1 | 0% | 1,282 | 2,039 | +59% | 0 | 0 | — |
case-12 | pass→pass | 12,684 | 14,524 | +15% | 1 | 1 | 0% | 1,900 | 3,395 | +79% | 0 | 0 | — |
case-13 | pass→fail | 17,270 | 17,579 | +2% | 1 | 1 | 0% | 2,511 | 3,861 | +54% | 0 | 0 | — |
case-14 | pass→pass | 17,077 | 23,910 | +40% | 1 | 1 | 0% | 2,744 | 5,304 | +93% | 0 | 0 | — |
case-15 | fail→fail | 17,049 | 22,237 | +30% | 1 | 1 | 0% | 2,662 | 4,765 | +79% | 0 | 0 | — |
case-16 | fail→pass | 20,040 | 22,354 | +12% | 1 | 1 | 0% | 3,056 | 4,574 | +50% | 0 | 0 | — |
case-17 | fail→pass | 12,869 | 15,243 | +18% | 1 | 1 | 0% | 1,912 | 3,459 | +81% | 0 | 0 | — |
case-18 | fail→pass | 15,575 | 13,233 | -15% | 1 | 1 | 0% | 2,366 | 3,121 | +32% | 0 | 0 | — |
case-19 | pass→pass | 28,018 | 10,700 | -62% | 1 | 1 | 0% | 2,976 | 2,703 | -9% | 0 | 0 | — |
case-20 | pass→pass | 23,126 | 26,452 | +14% | 1 | 1 | 0% | 3,495 | 5,273 | +51% | 0 | 0 | — |
case-21 | pass→pass | 24,816 | 34,263 | +38% | 1 | 1 | 0% | 4,077 | 6,597 | +62% | 0 | 0 | — |
case-22 | pass→pass | 23,917 | 33,960 | +42% | 1 | 1 | 0% | 3,924 | 6,705 | +71% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.